This research is designed to determine neuroanatomical and neurochemical correlates of addictive and aggressive/impulsive behavior in human subjects. The principal focus of these studies is the measurement and correlation of regional cerebral glucose metabolic activity, using positron emission tomography (PET), brain volumes using magnetic resonance imaging (MRI), cerebrospinal fluid metabolites, and measures of impulsive/aggressive behavior and excessive alcohol consumption. We collected full, volumetric T-1 weighted MR images using a 1.5 T scanner to measure intracranial volumes in 255 alcoholics (210 males and 91 females) and 157 healthy, non-alcoholic comparison subjects (80 males and 77 females). An automated segmentation program was used to divide the intracranial contents into CSF, gray and white matter (Human Brain Mapping, 5:194-205, 1997). When we measure brain volume we are measuring the combined effect of two processes: growth and degeneration. Growth determines maximum brain size achieved during life. Maximal brain growth can be estimated by intracranial volume (ICV) and since ICV remains constant throughout life, brain degeneration can be measured by the ratio of cerebral volume or gray matter or white matter volume to the remainder of the intracranial contents. Alcoholics show greater brain degeneration than non-alcoholics. Alcoholic women are more affected than alcoholic men. Alcoholics also show significantly greater brain shrinkage than controls by their mid to late twenties. In addition, alcoholics have smaller intracranial volumes than controls suggesting that pre-morbid differences in brain size may contribute to the risk for alcoholism. Despite the significant difference in intracranial volume brain, degeneration accounts for a greater amount of the difference in brain volume between alcoholics and controls than brain growth does. Similarly, presence or absence of co-morbid psychiatric disorder or other substance abuse does not affect brain shrinkage among alcoholics. Over the past year we have made several methodological advances in the automated measurement of brain volumes. An automated method for dividing the brain into right and left hemispheres was developed and validated. In addition, we have developed an automated method for measuring the volume of mesial and orbital frontal cortex, regions which are known to be involved in motivation and social behavior.